Selective determination of poly(styrene) and polyolefin microplastics in sandy beach sediments by gel permeation chromatography coupled with fluorescence detection

Marine Pollution Bulletin 136 (2018) 269–275

Contents lists available at ScienceDirect

Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul

Selective determination of poly(styrene) and polyolefin microplastics in sandy beach sediments by gel permeation chromatography coupled with fluorescence detection☆

T

Tarita Biver, Sabrina Bianchi, Maria Rita Carosi, Alessio Ceccarini, Andrea Corti, Enrico Manco, ⁎ Valter Castelvetro Department of Chemistry and Industrial Chemistry, University of Pisa, via G. Moruzzi 13, 56124 Pisa, Italy

A R T I C LE I N FO

A B S T R A C T

Keywords: Gel permeation chromatography Fluorescence Marine litter Microplastics Polystyrene Oxidized polyolefins

Microplastics generated by plastics waste degradation are ubiquitous in marine and freshwater basins, posing serious environmental concerns. Raman and FTIR spectroscopies, along with techniques such as pyrolysis-GC/ MS, are typically used for their identification. We present a procedure based on gel permeation chromatography (GPC) coupled with fluorescence detection for semi-quantitative selective determination of the most common microplastics found in marine shoreline sediments: poly(styrene) (PS) and partially degraded polyolefins (LDPEox). By operating the detector at either 260/280 or 370/420 nm excitation/emission wavelengths PS can be distinguished from LDPEox upon GPC separation. Semi-quantitative determination of microplastics contents is also possible: dichloromethane extracts of PS and LDPEox yield linear plots of fluorescence peak area vs concentration (0–5.0 mg/mL range) and were used as reference materials for quantification of the microplastics content in sand samples collected in the winter berm and dune sectors of a Tuscany beach in Italy.

1. Introduction The increasing attention towards the plastic pollution in water basins and their sediment systems is stimulating efforts aimed at improving sampling and characterization procedures. In this context, the pollution of sediments by microplastics has been investigated less extensively than that of the water column, with analytical methods still being developed and requiring standardisation and harmonisation (Sagawa et al., 2018; Van Cauwenberghe et al., 2015). Microplastics are minute plastics debris and particles with size ranging from few microns to a higher threshold varying, according to different researchers, from 500 μm up to 5 mm (Andrady, 2011). Their dispersion in the environment is a consequence of their release into wastewater as primary particles (e.g. synthetic fibers, microbeads), mainly from textile and personal care products, and of the fragmentation of larger plastic items caused by environmental degradation (photo-oxidative, hydrolytic) resulting in the generation of secondary particles. Commodity hydrocarbon polymers such as polyethylene (PE), polypropylene (PP), and polystyrene (PS) are those more likely to end up in shoreline rather than sea bottom sediments because of their low density (Andrady, 2017). For

these polymers environmental degradation occurs mainly through a photochemically assisted generation of oxidized functional groups (carbonyl, carboxyl, hydroxyl) followed by chain scissions and consequent reduction of the average molecular weight (Rabek, 1995). These processes weaken the integrity of plastic items that become brittle up to their powdery disintegration (Lambert and Wagner, 2016). It has been suggested that these polymers may produce a substantial contribution in the pollution of coastal sediments by microplastics since the processes of degradation and embrittlement of larger items proceed as they float at the sea/freshwater surface, and are accelerated once they are deposited ashore, where photo-oxidative, thermal and mechanical stresses are greater (Cooper and Corcoran, 2010; Ceccarini et al., 2018; Weinstein et al., 2016; Song et al., 2017). While mesh sizes smaller than 300 μm have also been used, (Dris et al., 2018; Setälä et al., 2016) sampling based on filtration of seawater or sieving of sediments has been mostly limited to the sampling to microplastics larger than 300–500 μm, although in the case of sediments sieving followed by density separation and filtration steps may allow collection and identification of microplastics down to 1–2 μm (Duis and Coors, 2016). However, isolation and characterization of individual small particles is

☆

T. Biver and A. Corti proposed and participated to the planning, experimentation and writing; S. Bianchi, M. R. Carosi and E. Manco carried out experimental work, A. Ceccarini supervised the experimental work and V. Castelvetro coordinated the research, provided funding and revised the paper. ⁎ Corresponding author. E-mail address: [email protected] (V. Castelvetro). https://doi.org/10.1016/j.marpolbul.2018.09.024 Received 1 August 2018; Received in revised form 11 September 2018; Accepted 14 September 2018 0025-326X/ © 2018 Elsevier Ltd. All rights reserved.

Marine Pollution Bulletin 136 (2018) 269–275

T. Biver et al.

dichloromethane (DCM) soluble fraction of a partially oxidized low density poly(ethylene) (LDPEox) obtained after catalytically enhanced oxidation (120 days thermal aging in air ventilated oven a 70 °C); ii) a commercial sample of expanded poly(styrene). As the environmental materials, DCM extracts previously collected from shoreline sand samples were analyzed (Ceccarini et al., 2018). Gel permeation chromatography (GPC) analyses were performed with an Agilent 1260 Infinity Binary LC instrument equipped with diode array (DAD VL+ 1260/G1315C) plus fluorescence (FLD 1260/G1321B) double detector, and two in series PLgel MIXED-E Mesopore columns (Polymer Laboratories) thermostated at 30 °C, using chloroform at 1.0 mL/min flow rate as the eluent. Ten polystyrene standards with molecular weights ranging from 800 to 300,000 g/mol−1 (Polymer Laboratories Ltd. and Varian, Inc.) were used for calibration. Both GPC detectors were calibrated for quantitative determination of the polymeric materials. For this purpose, five DCM solutions of PS (concentration range 0.02–4.15 mg/mL) and three DCM solutions of LDPEox (concentration range 0.54–5.61 mg/mL) were analyzed in triplicate by using the DAD detector at 261 nm and 243 nm, respectively, recording the average peak areas and relevant confidence interval. The same PS and LDPEox solutions were also used to calibrate the FLD detector once checked the most selective excitation and emission wavelengths for each polymeric material. Fluorescence spectra were separately recorded with a Perkin Elmer LS55 instrument, using reference materials and environmental samples in DCM, previously filtered on 0.2 μm PTFE membrane. Both excitation (wavelengths in the 200–775 nm range) and emission (200–800 nm range) spectra were recorded at 120 nm/min scan rate, adjusting the slits at 4 nm. It is worth pointing out that, although at a given wavelength the intensity of fluorescence emission F = ϕF(I-I0) (where ϕF is the quantum yield, I0 and I the incident and transmitted radiation, respectively) scales exponentially with the concentration C of the fluorophore along with Eq. (1),

not only impractical but also poorly representative of the diversity of this kind of pollution, also as a consequence of the contamination of microplastics with organic compounds and inorganic particles captured from the environment (Imhof et al., 2012). Chemical and enzymatic pre-treatments, including 30–35% hydrogen peroxide, (Nuelle et al., 2014) 30% HCl, and concentrated alkaline (e.g. NaOH) solutions, (Imhof et al., 2012; Rocha-Santos and Duarte, 2015) have been employed to remove organic contaminants from the microplastics in coastal sediment samples. Such aggressive chemical agents, however, may cause significant degradation or further alteration of the microplastics present in treated samples (Cole et al., 2014). Among the techniques used for the chemical identification of microplastics separated from sediment samples the most common are Fourier transformed infrared and Raman spectroscopies (Duis and Coors, 2016; Vianello et al., 2013) and, for smaller particles, micro-Raman and micro-FT-IR, (Song et al., 2015; Harrison et al., 2012; Kaeppler et al., 2015) the latter also associated with molecular imaging or focusing tools allowing the collection of spectra during visual inspection of the samples. Less practical for routine analysis but very effective for the chemical identification of plastics debris and their degradation products is pyrolysis coupled with gas-chromatography-mass spectrometry (Py-GC/MS) (Fries et al., 2013). Fabbri et al. (2000) used this technique to assess the contamination with poly(vinyl chloride) and other polymers in bottom sediments of a coastal lagoon, after isolation of the polymeric fractions by solvent extraction and re-precipitation in n-hexane. Direct analysis without polymer isolation suffers from the interference of inorganic components of the sediments affecting the degradation patterns of the polymeric materials during the analysis, (Fabbri et al., 1998) and from natural organic matter such as humic compounds whose thermal degradation produces the same congeners of synthetic polymers, particularly styrenic ones (van Loon et al., 1993; Audisio and Bertini, 1992). A two-step procedure based on thermal decomposition followed by absorption into a solid-phase device and subsequent GC/MS identification of the pyrolysis products has been proposed for the analysis of microplastics from environmental samples of complex composition (Duemichen et al., 2015; Duemichen et al., 2017). However, the same factors influencing the Py-GC/MS response may reduce the accuracy and reliability of the latter procedure. Several interfering factors may also affect the results of FT-IR and Raman spectroscopic analyses, including the morphology of single microplastics fragments, (Harrison et al., 2012) the presence of surface contaminants such as natural compounds, persistent organic pollutants (Ziccardi et al., 2016) or biofilms (Harrison et al., 2011; Lobelle and Cunliffe, 2011). Many of the cited methodologies comprise a separation step or the isolation and analysis of single microplastics fragments, limiting the minimum size of the fragment that can be reliably characterized and possibly excluding the extensively degraded ones, which may be more difficult to separate from the inorganic sediment or from biogenic debris. The present article is intended as a contribution to the improvement of the methodologies and techniques that, if used in a synergistic way, may provide more accurate information about level of contamination and fate of very small size microplastics and relevant degradation products in coastal sediments. In particular, gel permeation chromatography (GPC) equipped with refractive index, UV diode array and spectrofluorometric detectors has been used for qualitative and quantitative analysis of the polymer content in solvent extracts from marine sediments sampled in a sandy beach of northern Tuscany (Italy), which had been found to contain mainly low density polymers such as expanded polystyrene, polyethylene and polypropylene along with their partial degradation polymeric and oligomeric species (Ceccarini et al., 2018).

I = e−2.3 εbC I0

(1)

(where ε and b are the molar absorption coefficient and the optical path, respectively), the linear approximation I=I0(1–2.3εbC) that holds for Eq. (1) under conditions of negligible inner filter effect (low absorption coefficient at the excitation wavelength and low concentration) result in linearity of the response also for the fluorescence intensity, F = 2.3ϕFI0εbC. These are the conditions normally used for analytical applications and that are found to apply in the systems investigated here. 3. Results 3.1. Definition and optimization of instrumental parameters with reference materials In order to optimize the operating parameters for the spectrofluorometric detector to be used in the GPC analyses, 3D fluorescence spectra (x-axis: emission wavelength; y-axis: excitation wavelength; zaxis; intensity) of the two selected reference materials were recorded. The obtained fluorescence spectral map of polystyrene shows a maximum excitation at λex = 260 nm with a corresponding maximum emission centered at 335 nm (Fig. 1). These features of the fluorescence spectrum have been attributed to the emission from excimers formed between nearest neighbors pendant phenyl groups, (Torkelson et al., 1983; Vala et al., 1965; Torkelson et al., 1981; Hirayama, 1965) the fluorescence intensity being dependent on the molecular weight (MW) (Ishii et al., 1978). In the case of the LDPEox reference material the maximum emission can again be obtained at λex = 260 nm, resulting in two main emission bands centered at 358 and 375 nm, together with a bathochromically shifted, broad shoulder above 400 nm that appears as a single and more

2. Materials and methods Two polymeric materials representative of degraded poly(ethylene) and poly(styrene) were used as reference compounds: i) the 270

Marine Pollution Bulletin 136 (2018) 269–275

T. Biver et al.

400

exc

(nm)

380

B

180,0 160,0

360

140,0

340

120,0

320

100,0

300

80,00

280

60,00

260

40,00

240

20,00

220

200

200,0

A

150

F ( a .u .)

420

100

50

0,000

0

240 280 320 360 400 440 480 520 560 600

300

(nm) em

400

500

600

(nm)

Fig. 1. (A) 3D fluorescence spectral map the DCM extract of PS (x-axis = emission wavelength, y-axis = excitation wavelength, z-axis = intensity); (B) emission spectrum of a solution of PS in DCM with λexc = 260 nm.

400

exc

(nm)

380

200.0

A B

B

160.0

360

140.0

340

120.0

320

100.0

300

80.00

280

60.00

260

40.00

240

20.00

220

200

180.0

0.000

240 280 320 360 400 440 480 520 560 600 em

(nm)

150

F ( a .u .)

420

100

50

0

300

400

500

600

(nm)

Fig. 2. (A) 3D fluorescence spectral map of the DCM extract of LDPEox (x-axis = emission wavelength, y axis = excitation wavelength, z-axis = intensity); (B) emission spectra of a solution of LDPEox in DCM with λexc = 260 nm (solid line) and 370 nm (dashed line).

intense band upon excitation at λex = 370 nm. However, the 3D spectral map shows a secondary main emission centered at 420 nm upon excitation at 370 nm (Fig. 2). The presence of distinctive photoluminescence features in the PS and LDPEox samples allows using them as the reference materials for setting up the operating conditions of a GPC apparatus equipped with fluorescence detector, with the aim of assessing the suitability and accuracy of the technique for the selective semi-quantitative determination of naturally oxidized PS and polyolefins. In Fig. 3 are reported the GPC traces recorded with a DAD detector at 261 nm from the analysis of the DCM-soluble fractions of the two reference materials and of the DCM extracts obtained from beach sand samples collected in the winter berm and dune sectors (samples G3040011 and G3040012, respectively) of a touristic seashore site in north Tuscany, Italy (Ceccarini et al., 2018). The GPC trace of LDPEox in Fig. 3a presents a single broad and structured peak at high retention times (r.t.) associated with low-tomedium MW fractions (as expected, since high MW polyolefins are insoluble in DCM); on the other hand the GPC trace of the reference PS is characterized by a single peak at lower r.t. (its narrowness being most likely an artifact caused by a MW distribution covering a range close to the exclusion limit of the GPC columns), indicating the almost exclusive presence of high MW polymeric material. Indeed, the GPC traces from both sand DCM extracts (Fig. 3b) show the presence of both a narrow peak at low r.t. and a broad peak at high r.t., indicating the presence of both high and low-to-medium MW fractions, respectively. The UV spectra collected with the DAD detector in correspondence

with the narrow peak at low r.t. (structured absorption band with λmax = 262 nm and a secondary characteristic peak at λ = 269 nm) and with the broad peak at high r.t. (broad absorption band with λmax around 242 nm and a long tail extending up to nearly 400 nm) from the GPC fractionation of the two sand extracts matched quite well those recorded under the same conditions from the reference PS and LDPEox, respectively (Fig. 4). In order to check the accuracy of the DAD detector for the evaluation of the concentrations of the two polymer types (PS and oxidized polyolefins) in real samples, two calibration curves were obtained by running GPC analysis of PS and LDPEox DCM solutions at different concentrations. The analyses, performed in triplicate for each concentration of the reference materials, gave a strictly linear relationship between concentration and DAD peak area within the explored range of 25–5500 mg/L for PS and of 720–7400 mg/L for LDPEox. The GPC runs with the same reference PS and LDPEox solutions and with the two beach sand DCM extracts were then replicated in four subsequent experiments by setting up the fluorescence detector (FLD) at four different excitation/emission wavelengths combinations: 260/ 280 nm, 260/335 nm, 370/395 nm, and 370/420 nm. Representative GPC traces obtained from the same sample but with different FLD setup (excitation/emission wavelengths at 260/280 nm and 370/420 nm, respectively) are shown in Fig. 5. In accordance with the fluorescence spectral map of the reference materials (Figs. 1 and 2), when the fluorescence detector was set at 260/280 nm excitation/emission wavelengths the fluorescence contribution of LDPEox (DCM-soluble fraction) was found to be negligible, 271

Marine Pollution Bulletin 136 (2018) 269–275

T. Biver et al.

Fig. 3. GPC traces recorded with DAD detector at λ = 261 nm: (a) DCM-soluble fraction of LDPEox and from PS (whole sample) reference materials; (b) DCM extracts of the sand samples from the winter berm (sample G3040011) and dune (sample G3040012) sectors.

as opposed to the strong fluorescence response associated with PS materials (Fig. 5a–b). Such a clear-cut discrimination of the two polymeric materials was obtained neither by recording the emissions at 335 nm with the same excitation wavelength, nor with the FLD set at 370/385 nm excitation/emission. On the other hand, with the FLD set at 370/420 nm excitation/emission the fluorescent response of PS-like materials becomes negligible, while LDPEox and, more generally, oxidized polyolefins exhibit a residual fluorescence sufficient for their selective quantification (Fig. 5c–d). Based on the above results, the 260/280 and 370/420 nm combinations of excitation/emission wavelengths can be considered as suitable to distinguish and recognize polystyrene-like materials from degraded and oxidized polyolefins within the same sample along with GPC separation. To assess the accuracy of the detection also for the quantitative analysis of these two classes of polymeric materials, and more specifically for their determination as environmental contaminants extracted as DCM soluble fraction from coastal sand sediments, two distinct calibration curves were obtained by analyzing DCM solutions of the reference PS (5.51–0.033 mg/mL range) and DCM extracts of the reference LDPEox (4.33–0.57 mg/mL range) with the FLD set at 260/ 280 nm and 370/420 nm excitation/emission, respectively. In Fig. 6 are reported the linear fits of the calibration curves obtained by considering the GPC peak areas from the GPC/FLD traces. The experimental (weighed extracts) and calculated data from the quantitative analysis of the two environmental samples G3040011

(winter berm) and G3040012 (dune) are reported in Table 1, where the calculated data are based on DAD and FLD detection of the GPC eluates, respectively, and POox indicates the DCM-soluble fraction of degraded polyolefins in general, including LDPE, HDPE (high density polyethylene), PP, and similar semicrystalline olefin copolymers. The main source of the mismatch between the weighed total amounts of extracted microplastic material and the sum of PS and POox as calculated from the FLD peak areas based on the relevant calibration curves (e.g. 18.7 g vs. 8.4 + 3.3 = 11.7 g, respectively, for the winter berm sample) could be ascribed to the heterogeneous nature of the POox fraction originated from the environmental pool. As a means to clarify this issue, as discussed later, in Fig. 7 are reported the GPC traces of the reference LDPEox recorded with three different detectors, namely refractive index (RI), providing a signal intensity scaling linearly with concentration by mass, UV and FLD, the intensity of their response being proportional to the concentration of the absorbing/emitting chromophores/fluorophores. The peak intensities are normalized for easier comparison. 4. Discussion Photoluminescence of polyolefins is somewhat controversial, (Allen et al., 2000) having been associated with the presence of oxidized “impurities” such as carbonyl end groups, (Charlesby and Partridge, 1965; Allen et al., 1977a) α,β-unsaturated carbonyl (enone and dione types) (Allen et al., 1977b) and dicarbonyl species. Photoluminescence

140

160

(a)

120

140

(b)

120

100

100

AU

AU

80

80

60 60

40

40

20

20

0

0

230

250

270

290

310

330

350

370

wl (nm)

230 240 250 260 270 280 290 300 310 320 330 340 350 360 370

wl (nm)

Fig. 4. DAD UV spectra of GPC elution fractions of DCM extracts from test samples and reference materials recorded at 10.5 min (a) and 16.5 min (b) retention time. Black lines: reference PS (A) and LDPEox (B) polymeric material; red lines: winter berm sample; blue lines dune sample. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 272

Marine Pollution Bulletin 136 (2018) 269–275

T. Biver et al.

Fig. 5. GPC traces recorded with different FLD set ups: with 260/280 nm excitation/emission wavelengths from LDPEox and PS reference materials (a) and from the DCM extracts of the beach sand samples from the winter berm and dune sectors (b); with FLD set up at 370/420 nm excitation/emission wavelength from LDPEox and PS reference materials (c) and from the DCM extracts of winter berm and dune samples (d).

ascribed to the loss of fluorescent additives (e.g. antioxidants) has been observed, followed by a substantial increase closely related with the evolution of degradation processes (Grabmayer et al., 2014a). Such post-oxidative luminescence has been generally attributed to the initial formation of α,β-unsaturated carbonyl and similarly oxygenated groups, followed by the development of more extended conjugated systems such as short polyene sequences that may or may not be conjugated to carbonyl groups; (Jacques and Poller, 1993; Osawa and Kuroda, 1986) the latter are typically excited at longer wavelengths (between 300 and 400 nm) with ensuing fluorescence at wavelengths exceeding 500 nm (Grabmayer et al., 2014a; Grabmayer et al., 2014b). Owing to these properties, selective detection of PS and LDPEox photoluminescence is possible, because the former would not be excited (and thus would not exhibit any photoluminescence) upon excitation in the visible range of wavelengths, as opposed to the latter. The complementary spectral response for the two polymer species is highlighted in Fig. 5a,c, in which the GPC traces for the DCM extracts of the two reference materials are recorded with the FLD detector set at either 260/280 (for the selective detection of PS) or 370/420 nm (for LDPEox) excitation/emission wavelengths. The GPC traces recorded from the DCM extracts of the beach samples, Fig. 5b,d, show in both cases the presence of a high and a low-tomedium MW fraction. Despite the inherently complex composition of the samples, the UV spectra collected with the DAD detector in correspondence with the narrow GPC peak at low r.t. and with the broad peak at high r.t. were found to match with those of the reference PS and LDPEox, respectively (Fig. 4). This suggests that the chosen reference materials are appropriate to describe the DCM extracts from the sandy

Fig. 6. GPC calibration curves based on FLD detector. Left: reference PS (λexc = 260 nm; λem = 280 nm); right: reference LDPEox (λexc = 370 nm; λem = 420 nm).

and chemiluminescence properties of oxidized polyolefins have already been exploited for quantitative analyses because of their higher sensitivity with respect to e.g. FT-IR spectroscopy; the kinetics of thermoand photo-oxidative degradation of polyolefins could thus be studied from the very onset, when only few oxidized groups are present (Jacques and C, 1993; Steffen et al., 2016; Grabmayer et al., 2014a; Tiemblo et al., 1999). In most cases, an initial decrease in the emission, 273

Marine Pollution Bulletin 136 (2018) 269–275

T. Biver et al.

Table 1 Content and composition of the DCM extracts from environmental sand samples calculated using the DAD and FLD calibration curves based on PS and LDPEox reference materials. Sampling sector

Total extractsa (mg/kg)

Sample

Dune Winter berm

G3040012 G3040011

High MW PSb (mg/kg)

327.0 18.7

Total PSc (mg/kg)

POoxd (mg/kg)

DAD

FLD260/280

FLD260/280

FLD370/420

57.0 ± 7.6 6.1 ± 1.5

66.1 ± 4.1 5.7 ± 0.5

104.3 ± 7.0 8.4 ± 1.0

76.9 3.3

a

Weighed amount of polymeric material extracted with DCM from 1 kg sand; Calculated from the lower r.t. GPC peak area (see reference (Ceccarini et al., 2018)); c Calculated from the combined low r.t. (high MW fraction) and high r.t. (low MW fraction) GPC peak areas using the FLD260/280 detector and related PS calibration curve. d Calculated from the calibration curve obtained using LDPEox as the reference material. b

mass, in agreement with the observed mismatch between the response of FLD and RI detectors. Therefore, depending on the actual composition of a given environmental POox sample, the calibration based on LDPEox may result in either underestimation (for highly degraded microplastics with a larger low-MW fraction) or overestimation of the actual amount of POox. However, since extensive degradation is required for polyolefin fragmentation and pulverization, the structural features of different polyolefins once turned into microplastics are likely to be similar (Chiellini et al., 2006).

beach samples, irrespective of the fact that LDPEox is actually a heterogeneous mixture of variously oxidized medium-to-low MW LDPE degradation products of a single polyolefin type. Concerning the PS content, both FLD and DAD detectors allow quantitative analysis for the high MW fraction, as this is nearly pure PS since even partially oxidized polyolefins are not soluble DCM at molecular weights exceeding a few thousand Daltons. On the other hand, while DAD does not allow discrimination between PS and POox, by setting the FLD detector at 260/280 nm excitation/emission it is possible to selectively perform quantitative analysis of the PS content in the low MW fraction containing also POox, and thus the total amount of PS can be determined. Instead, the figures calculated from the GPC peak area with FLD set for POox detection and based on the calibration with LDPEox appear to largely underestimate the actual content in the environmental extracts as determined by gravimetry (see Table 1). Such a mismatch may be at least partially ascribed to the presence of low MW hydrocarbon species such as natural waxes and heavy aliphatic hydrocarbon pollutants (e.g. from oil spillage). However, an additional source of inaccuracy may be ascribed to the intrinsic limitations of the spectroscopic detection, associated with calibration based on a specific LDPEox reference material, for the quantitative determination of the DCM-soluble fraction of the variable pool of POox from sampled microplastics. The GPC traces in Fig. 7 clearly show that the two spectroscopic detectors, and FLD in particular, underestimate the amount of higher MW fraction and overestimate the lower MW one as compared to the RI detector. Such discrepancy can be ascribed to the differences in fluorophores (photoluminescent moieties) molar concentrations per unit mass of POox as a function of MW, a result of photooxidative chain scission occurring largely by beta elimination with formation of carbonyl groups at the chain end, which may then further react to yield fluorescent conjugated moieties (Jacques and C, 1993). Thus, the shorter the residual chain fragment, the higher the molar concentration of fluorophores per unit

GPC chromatography with fluorescence detection (FLD) allows selective determination of hydrocarbon-based microplastics pollutants sampled from marine and freshwater coastal sediments. Linear FLD response with concentration was found for both PS and LDPEox reference materials used as calibration standards. Accurate direct determination of PS can be achieved, while for polyolefins the analysis is semi-quantitative, being limited to the DCM-soluble, highly degraded polyolefin fraction. In the latter case a lower accuracy may result from the intrinsic structural variability of the POox pool and from the environmental contamination by hydrocarbons pollutants (e.g. from oil spillage). Compared to the most common analytical procedures based on particle counting and identification, the proposed method based on solvent extraction and GPC-FLD analysis provides no information on the size distribution of the microplastics; besides, different polyolefins cannot be distinguished. Nevertheless, the simple and straightforward method proposed here allows performing qualitative and semi-quantitative determination of the microplastics from marine sediments, preferably in combination with gravimetric determination of the whole DCM soluble fraction.

y= 6.93x - 0.01 R2 = 0.9999

FLD370/420 peak area (a.u)

FLD260/280 peak area (a.u)

40

5. Conclusions

30 20 10 0 0.0

1.0

2.0

3.0 4.0 mg/mL

5.0

y = 7,08x + 0,28 R² = 0,9982

30 20 10 0 0.0

6.0

1.0

2.0 3.0 mg/mL

4.0

5.0

Fig. 7. Normalized GPC traces of LDPEox as recorded with refractive index (RI), DAD (UV absorption at λ = 260 nm) and FLD (λexc = 370 nm; λem = 420 nm) detectors.

274

Marine Pollution Bulletin 136 (2018) 269–275

T. Biver et al.

Finally, it is worth pointing out that the onset of pulverization of a macroscopic polyolefin item, resulting in the generation of secondary microplastics, occurs when thorough oxidative degradation has already been achieved. Thus the DCM-soluble fraction is likely to be the main one in polyolefin microplastic particles and its fluorescence response is not expected to be significantly affected by the type of original polyolefin material.

Raeder, B., Buchberger, W., Lang, R.W., 2014b. Characterization of the aging behavior of polyethylene by photoluminescence spectroscopy. Polym. Degrad. Stab. 107, 28–36. Harrison, J.P., Sapp, M., Schratzberger, M., Osborn, A.M., 2011. Interactions between microorganisms and marine microplastics: a call for research. Mar. Technol. Soc. J. 45, 12–20. Harrison, J.P., Ojeda, J.J., Romero-Gonzalez, M.E., 2012. The applicability of reflectance micro-Fourier-transform infrared spectroscopy for the detection of synthetic microplastics in marine sediments. Sci. Total Environ. 416, 455–463. Hirayama, F., 1965. Intramolecular excimer formation. I. Diphenyl and triphenyl alkanes. J. Chem. Phys. 42, 3163–3171. Imhof, H.K., Schmid, J., Niessner, R., Ivleva, N.P., Laforsch, C., 2012. A novel, highly efficient method for the separation and quantification of plastic particles in sediments of aquatic environments. Limnol. Oceanogr. Methods 10, 524–537. Ishii, T., Handa, T., Matsunaga, S., 1978. Effect of molecular weight on excimer formation of polystyrenes in solution. Macromolecules 11, 40–46. Jacques, P.P.L., C, P.R., 1993. Fluorescence of polyolefins-1. Effect of thermal degradation on fluorescent excitation and emission spectra. Eur. Polym. J. 29, 75–81. Jacques, P.P.L., Poller, R.C., 1993. Fluorescence of polyolefins—2. Use of model compounds to identify fluorescent species in thermally degraded polymers. Eur. Polym. J. 29, 83–89. Kaeppler, A., Windrich, F., Laeder, M.G.J., Malanin, M., Fischer, D., Labrenz, M., Eichhorn, K.-J., Voit, B., 2015. Identification of microplastics by FTIR and Raman microscopy: a novel silicon filter substrate opens the important spectral range below 1300 cm(−1) for FTIR transmission measurements. Anal. Bioanal. Chem. 407, 6791–6801. Lambert, S., Wagner, M., 2016. Formation of microscopic particles during the degradation of different polymers. Chemosphere 161, 510–517. Lobelle, D., Cunliffe, M., 2011. Early microbial biofilm formation on marine plastic debris. Mar. Pollut. Bull. 62, 197–200. Nuelle, M.-T., Dekiff, J.H., Remy, D., Fries, E., 2014. A new analytical approach for monitoring microplastics in marine sediments. Environ. Pollut. 184, 161–169. Osawa, Z., Kuroda, H., 1986. Differences in polyene formation between polyethylene and polypropylene during photo-irradiation. Polym. Photochem. 7, 231–236. Rabek, J.F., 1995. Polymer Photodegradation. Mechanisms and Experimental Methods. Chapman & Hill, London, UK. Rocha-Santos, T., Duarte, A.C., 2015. A critical overview of the analytical approaches to the occurrence, the fate and the behavior of microplastics in the environment. Trends Anal. Chem. 65, 47–53. Sagawa, N., Kawaai, K., Hinata, H., 2018. Abundance and size of microplastics in a coastal sea: comparison among bottom sediment, beach sediment, and surface water. Mar. Pollut. Bull. 133, 532–542. Setälä, O., Magnusson, K., Lehtiniemi, M., Norén, F., 2016. Distribution and abundance of surface water microlitter in the Baltic Sea: A comparison of two sampling methods. Mar. Pollut. Bull. 110, 177–183. Song, Y.K., Hong, S.H., Jang, M., Han, G.M., Rani, M., Lee, J., Shim, W.J., 2015. A comparison of microscopic and spectroscopic identification methods for analysis of microplastics in environmental samples. Mar. Pollut. Bull. 93, 202–209. Song, Y.K., Hong, S.H., Jang, M., Han, Gi M., Seung, W.J., Shim, W.J., 2017. Combined effects of UV exposure duration and mechanical abrasion on microplastic fragmentation by polymer type. Environ. Sci. Technol. 51, 4368–4376. Steffen, R., Wallner, G., Rekstad, J., Röder, B., 2016. General characteristics of photoluminescence from dry heat aged polymeric materials. Polym. Degrad. Stab. 134, 49–59. Tiemblo, P., Gòmez-Elvira, J.M., Teyssedre, G., Massine, F., Laurent, C., 1999. Chemiluminescence spectral evolution along the thermal oxidation of isotactic polypropylene. Polym. Degrad. Stab. 65, 113–121. Torkelson, J.M., Lipsky, S., Tirrell, M., 1981. Polystyrene fluorescence: effects of molecular weight in various solvents. Macromolecules 14, 1601–1603. Torkelson, J.M., Lipsky, S., Tirrell, M., Tirrell, D.A., 1983. Fluorescence and absorbance of polystyrene in dilute and semidilute solutions. Macromolecules 16, 326–330. Vala, M.T., Haebig, J., Rice, S.A., 1965. Experimental study of luminescence and excitation trapping in vinyl polymers, paracyclophanes, and related compounds. J. Chem. Phys. 43, 886–897. Van Cauwenberghe, L., Devriese, L., Galgani, F., Robbens, J., Janssen, C.R., 2015. Microplastics in sediments: a review of techniques, occurrence and effects. Mar. Environ. Res. 111, 5–17. van Loon, W.M.G.M., Boon, J.J., de Groot, B., 1993. Quantitative determination of macromolecular chlorolignosulfonic acids in water by pyrolysis-gas chromatography/mass spectrometry with single ion monitoring. Anal. Chem. 65, 1728–1735. Vianello, A., Boldrin, A., Guerriero, P., Moschino, V., Rella, R., Sturaro, A., Da Ros, L., 2013. Microplastic particles in sediments of Lagoon of Venice, Italy: first observations on occurrence, spatial patterns and identification. Estuar. Coast. Shelf Sci. 130, 54–61. Weinstein, J.E., Crocker, B.K., Gray, A.D., 2016. From macroplastic to microplastic: degradation of high-density polyethylene, polypropylene, and polystyrene in a salt marsh habitat. Environ. Toxicol. Chem. 35, 1632–1640. Ziccardi, L.M., Edgington, A., Hentz, K., Kulacki, K.J., Driscoll, S.K., 2016. Microplastics as vectors for bioaccumulation of hydrophobic organic chemicals in the marine environment: a state-of-the-science review. Environ. Toxicol. Chem. 35, 1667–1676.

Acknowledgements This work was supported by the University of Pisa [PRA 2017_17 project]. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.marpolbul.2018.09.024. References Allen, N.S., Homer, J., McKellar, J.F., 1977a. Origin and role of the luminescent species in the photo-oxidation of commercial polypropylene. J. Appl. Polym. Sci. 21, 2261–2267. Allen, N.S., Homer, J., McKellar, J.F., Wood, D.G.M., 1977b. Identification of the luminescent species in low-density polyethylene. J. Appl. Polym. Sci. 21, 3147–3152. Allen, N.S., Edge, M., Holdsworth, D., Rahman, A., Catalina, F., Fontan, E., Escalona, A.M., Sibon, F.F., 2000. Ageing and spectroscopic properties of polyethylenes: comparison with metallocene polymer. Polym. Degrad. Stab. 67, 57–67. Andrady, A.L., 2011. Microplastics in the marine environment. Mar. Pollut. Bull. 62, 1596–1605. Andrady, A.L., 2017. The plastic in microplastics: a review. Mar. Pollut. Bull. 119, 12–22. Audisio, G., Bertini, F., 1992. Molecular weight and pyrolysis products distribution of polymers I. Polystyrene. J. Anal. Appl. Pyrolysis 24, 61–74. Ceccarini, A., Corti, A., Erba, F., Modugno, F., La Nasa, J., Bianchi, S., Castelvetro, V., 2018. From macroplastic to microplastic: degradation of high-density polyethylene, polypropylene, and polystyrene in a salt marsh habitat. Environ. Sci. Technol. 52, 5634–5643. Charlesby, A., Partridge, R.H., 1965. The identification of luminescence centres in polyethylene and other polymers. Proc. R. Soc. A 283, 312–328. Chiellini, E., Corti, A., D'Antone, S., Baciu, R., 2006. Oxo-biodegradable carbon backbone polymers – oxidative degradation of polyethylene under accelerated test conditions. Polym. Degrad. Stab. 91, 2739–2747. Cole, M., Webb, H., Lindeque, P.K., Fileman, E.S., Halsband, C., Galloway, T.S., 2014. Isolation of microplastics in biota-rich seawater samples and marine organisms. Sci. Rep. 4, 4528. Cooper, D.A., Corcoran, P.L., 2010. Effects of mechanical and chemical processes on the degradation of plastic beach debris on the island of Kauai, Hawaii. Mar. Pollut. Bull. 60, 650–654. Dris, R., Gasperi, J., Rocher, Vincent., Tassin, B., 2018. Synthetic and non-synthetic anthropogenic fibers in a river under the impact of Paris megacity: sampling methodological aspects and flux estimations. Sci. Total Environ. 618, 157–164. Duemichen, E., Barthel, A.-K., Braun, U., Bannick, C.G., Brand, K., Jekel, M., Senz, R., 2015. Analysis of polyethylene microplastics in environmental samples, using a thermal decomposition method. Water Res. 85, 451–457. Duemichen, E., Eisentraut, P., Bannick, C.G., Barthel, A.-K., Senz, R., Braun, U., 2017. Fast identification of microplastics in complex environmental samples by a thermal degradation method. Chemosphere 174, 572–584. Duis, K., Coors, A., 2016. Microplastics in the aquatic and terrestrial environment: sources (with a specific focus on personal care products), fate and effects. Environ. Sci. Eur. 28, 2. Fabbri, D., Trombini, C., Vassura, I., 1998. Analysis of polystyrene in polluted sediments by pyrolysis-gas chromatography-mass spectrometry. J. Chromatogr. Sci. 36, 600–604. Fabbri, D., Tartari, D., Trombini, C., 2000. Analysis of poly(vinyl chloride) and other polymers in sediments and suspended matter of a coastal lagoon by pyrolysis-gas chromatography-mass spectrometry. Anal. Chim. Acta 413, 3–11. Fries, E., Dekiff, J.H., Willmeyer, J., Nuelle, M.-T., Ebert, M., Remy, D., 2013. Identification of polymer types and additives in marine microplastic particles using pyrolysis-GC/MS and scanning electron microscopy. Environ. Sci.: Processes Impacts 15, 1949–1956. Grabmayer, K., Wallner, G.M., Beissmann, S., Braun, U., Steffen, R., Nitsche, D., Raeder, B., Buchberger, W., Lang, R.W., 2014a. Accelerated aging of polyethylene materials at high oxygen pressure characterized by photoluminescence spectroscopy and established aging characterization methods. Polym. Degrad. Stab. 109, 40–49. Grabmayer, K., Wallner, G.M., Beissmann, S., Schlothauer, J., Steffen, R., Nitsche, D.,

275